{
  "id": 198530,
  "title": "Which way is the most appropriate for making submission?",
  "url": "/competitions/cassava-leaf-disease-classification/discussion/198530",
  "author_name": "",
  "post_date": "2020-11-21T17:17:58.060150600Z",
  "votes": 5,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi, </p>\n<p>For submitting the notebook, should the code include the training part as well? Or can we just train the model offline and load inside from the inference notebook and submit that. How should be the process? Thanks.</p>",
  "messages": [
    {
      "id": "1086432",
      "postDate": "11/21/2020 17:17:58",
      "content": "<p>Hi, </p>\n<p>For submitting the notebook, should the code include the training part as well? Or can we just train the model offline and load inside from the inference notebook and submit that. How should be the process? Thanks.</p>",
      "rawMarkdown": "Hi, \n\nFor submitting the notebook, should the code include the training part as well? Or can we just train the model offline and load inside from the inference notebook and submit that. How should be the process? Thanks.",
      "votes": null
    },
    {
      "id": "1086498",
      "postDate": "11/21/2020 18:21:13",
      "content": "<p><a href=\"https://www.kaggle.com/snnclsr\" target=\"_blank\">@snnclsr</a> You can train the model offline and load the inference notebook .</p>",
      "rawMarkdown": "snnclsr You can train the model offline and load the inference notebook .",
      "votes": null
    },
    {
      "id": "1086505",
      "postDate": "11/21/2020 18:31:15",
      "content": "<p>Hi, Thanks for your answer. I will go with that option. </p>",
      "rawMarkdown": "Hi, Thanks for your answer. I will go with that option.",
      "votes": null
    },
    {
      "id": "1132317",
      "postDate": "12/30/2020 09:59:47",
      "content": "<p>Please what exactly do you mean by train the model offline and inference notebook?<br>\nAm I supposed to use one notebook for training and another notebook for prediction and submission?<br>\nIf so, should I delete the previous notebook for training?<br>\nAnd what do you mean by train the model offline?</p>",
      "rawMarkdown": "Please what exactly do you mean by train the model offline and inference notebook?\nAm I supposed to use one notebook for training and another notebook for prediction and submission?\nIf so, should I delete the previous notebook for training?\nAnd what do you mean by train the model offline?",
      "votes": null
    },
    {
      "id": "1132325",
      "postDate": "12/30/2020 10:09:12",
      "content": "<p>Hi, since this is a code competition, you must submit your work via the notebooks. Training offline means that if you have a local gpu, you can train your model \"offline\" which doesnt restrict you to use kaggle or colab notebooks for training. Then you can upload your models to kaggle as a dataset, and use that in your inference notebook. If you have a one train and one inference notebook which is perfectly okay, you can just continue to use them. This competition expects you to submit your inference notebook. I hope it's more clear.</p>",
      "rawMarkdown": "Hi, since this is a code competition, you must submit your work via the notebooks. Training offline means that if you have a local gpu, you can train your model \"offline\" which doesnt restrict you to use kaggle or colab notebooks for training. Then you can upload your models to kaggle as a dataset, and use that in your inference notebook. If you have a one train and one inference notebook which is perfectly okay, you can just continue to use them. This competition expects you to submit your inference notebook. I hope it's more clear.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1086498,
      "author_name": "usharengaraju",
      "author_url": "",
      "post_date": "11/21/2020 18:21:13",
      "content": "<p><a href=\"https://www.kaggle.com/snnclsr\" target=\"_blank\">@snnclsr</a> You can train the model offline and load the inference notebook .</p>",
      "votes": null,
      "replies": [
        {
          "id": 1086505,
          "author_name": "snnclsr",
          "author_url": "",
          "post_date": "11/21/2020 18:31:15",
          "content": "<p>Hi, Thanks for your answer. I will go with that option. </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1132317,
          "author_name": "black9t",
          "author_url": "",
          "post_date": "12/30/2020 09:59:47",
          "content": "<p>Please what exactly do you mean by train the model offline and inference notebook?<br>\nAm I supposed to use one notebook for training and another notebook for prediction and submission?<br>\nIf so, should I delete the previous notebook for training?<br>\nAnd what do you mean by train the model offline?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1132325,
          "author_name": "snnclsr",
          "author_url": "",
          "post_date": "12/30/2020 10:09:12",
          "content": "<p>Hi, since this is a code competition, you must submit your work via the notebooks. Training offline means that if you have a local gpu, you can train your model \"offline\" which doesnt restrict you to use kaggle or colab notebooks for training. Then you can upload your models to kaggle as a dataset, and use that in your inference notebook. If you have a one train and one inference notebook which is perfectly okay, you can just continue to use them. This competition expects you to submit your inference notebook. I hope it's more clear.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1086432": "Hi, \n\nFor submitting the notebook, should the code include the training part as well? Or can we just train the model offline and load inside from the inference notebook and submit that. How should be the process? Thanks.",
    "1086498": "snnclsr You can train the model offline and load the inference notebook .",
    "1086505": "Hi, Thanks for your answer. I will go with that option.",
    "1132317": "Please what exactly do you mean by train the model offline and inference notebook?\nAm I supposed to use one notebook for training and another notebook for prediction and submission?\nIf so, should I delete the previous notebook for training?\nAnd what do you mean by train the model offline?",
    "1132325": "Hi, since this is a code competition, you must submit your work via the notebooks. Training offline means that if you have a local gpu, you can train your model \"offline\" which doesnt restrict you to use kaggle or colab notebooks for training. Then you can upload your models to kaggle as a dataset, and use that in your inference notebook. If you have a one train and one inference notebook which is perfectly okay, you can just continue to use them. This competition expects you to submit your inference notebook. I hope it's more clear."
  },
  "source": "meta"
}